Modelling THM formation potential based on the character of organics - in catchments and drinking water sources

Abstract

Problems caused by dissolved organic matter (DOM) present in drinking water include its constituent compounds being precursors in the formation of disinfection by-products that pose risks to human health. The prediction of trihalomethanes (THM) formation is highly challenging due to a wide range of impacting factors, including the concentration and character of DOM, pH, temperature and the presence of halide anions (Cl-, Br-) in drinking water. The aim of the research reported here was to develop model(s) for prediction of THM formation potential (THMFP) including the four compounds, chloroform, bromodichloromethane, chlorodibromomethane and bromoform, based on the character of precursor organics sourced from discrete catchments of a drinking water reservoir. Water samples were collected from six discrete zero-order catchments (ZOCs) of the Myponga reservoircatchment, South Australia under three land management practices (Australian native vegetation, pine plantation, grasslands) with varying soil textures. Water samples were also collected from the main stream (Myponga River) and the Myponga Reservoir. Water samples from the ZOCs were collected at both the surface and subsurface (~60 cm depth) and were treated by alum coagulation under laboratory conditions to simulate conventional treatment for removal of organic compounds. Samples were analyzed for dissolved organic carbon (DOC) and UV-visible absorbance to measure and characterize the DOM. Raw and alum treated waters were then tested for THMFP under standardized laboratory conditions. Mathematical models were developed to predict THMFP and abundances of constituent compounds from the DOM concentration and character, and bromide (Br) ion concentration (to ~0.5 mg/L), based on regression analysis. A linear model (R2 = 0.97, T-test = 0.92 and standard error: SE = 32.9 μg/L) was developed to describe the relationship between THMFP and the character of DOM in terms of aromatic and non-aromatic compounds. Similarly, a linear model (R2 = 0.95, T-test = 0.89 and SE = 33.9 μg/L) was fitted to describe the relationship between chloroform (CHCl3) concentration and DOM character. For modelling of the formation of constituent THM compounds, the percentage formation (%) of CHCl3 was predicted by using the same independent variables (DOC, UV absorbance at 254 nm (UV254) and Br), (R2 = 0.85, T-test = 0.99 and SE = 8.8%). Using the relative abundance of chloroform to THMFP, the percentages of bromodichloromethane and chlorodibromomethane were estimated using a power function equation (R2 = 0.99) for bromodichloromethane and an exponential function equation (R2 = 0.98) for chlorodibromomethane. From previous models, the concentration of each individual THM compound can be estimated. From this study, it was found that data of the character of DOM and Br present in waters of catchments with discrete land use can be related to THMFP and abundances of constituent compounds.